Research map / Large language models / Retrieval and knowledge
PharmaQA: Prompt-Based Molecular Representation Learning via Pharmacophore-Oriented Question Answering
Chengwei Ai, Qiaozhen Meng, Mengwei Sun, Ruihan Dong, Hongpeng Yang, Shiqiang Ma, Xiaoyi Liu, Cheng Liang, Fei Guo
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add learned pharmacophore QA prompts to the graph encoder (ablation) | PharmaQA · MoleculeNet classification (6 tasks, average) | AUC: 0.825 -> 0.842 | Improved |
| add learned pharmacophore QA prompts to the graph encoder (ablation) | PharmaQA · MoleculeNet regression (3 tasks, average) | RMSE: 1.175 -> 1.039 | Improved |
| use meaningful pharmacophore prompts instead of an irrelevant noise sentence (ablation) | PharmaQA · MoleculeNet classification (6 tasks, average) | AUC: 0.791 -> 0.842 | Improved |
| use learned prompts instead of static RDKit pharmacophore features (ablation) | PharmaQA · MoleculeNet regression (3 tasks, average) | RMSE: 1.251 -> 1.039 | Improved |
| static RDKit pharmacophore features instead of no prompt (ablation) | graph-only baseline · MoleculeNet classification (6 tasks, average) | AUC: 0.825 -> 0.803 | Hurt |
| noise (irrelevant sentence) prompt instead of no prompt (ablation) | graph-only baseline · MoleculeNet regression (3 tasks, average) | RMSE: 1.175 -> 1.207 | Hurt |
| PharmaQA vs KPGT on ToxCast | PharmaQA · MoleculeNet ToxCast (scaffold split) | AUC: 0.746 -> 0.746 | No clear effect |
| PharmaQA vs KPGT on ligand affinity datasets | PharmaQA · FGFR1, HPK1, VIM-1 ligand datasets | Spearman and Pearson correlation: FGFR1 +0.038 Spearman, +0.020 Pearson | Improved |
Closest papers by meaning
- Knowledge Enhanced Representation Learning for Drug Discovery (AAAI 2024 · similarity 0.87)
- PharmacoMatch: Efficient 3D Pharmacophore Screening via Neural Subgraph Matching (ICLR 2025 · similarity 0.87)
- MolTailor: Tailoring Chemical Molecular Representation to Specific Tasks via Text Prompts (AAAI 2024 · similarity 0.86)
- Drug-TTA: Test-Time Adaptation for Drug Virtual Screening via Multi-task Meta-Auxiliary Learning (ICML 2025 · similarity 0.84)
- Mol-LLaMA: Towards General Understanding of Molecules in Large Molecular Language Model (NeurIPS 2025 · similarity 0.84)
- Atomas: Hierarchical Adaptive Alignment on Molecule-Text for Unified Molecule Understanding and Generation (ICLR 2025 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
